High Level Mission Assignment Optimization

Our research deals with Weapon-Target Assignment (WTA) problem which is a combinatorial optimization one that known to be NP-complete. The WTA is an optimization problem focus on setting the best assignment of weapons to targets, to minimizing the total expected value of the surviving targets. In this paper, we compared different methods of Z3 and Simulated Annealing (SA) to solve the WTA problem and suggest a novel algorithm based on Deep Q Networks (DQN) method. The main advantage of the DQN algorithm since we can learn in advance what is the optimal action for every space in short amount of time. Moreover, in real time the actual assignment of weapon can be done in a lot shorter amount of time then in the SA algorithm and the Z3 as can be seen in the results presented in this paper.